Archaeology and the Biblical Narrative: The Case of the United Monarchy

Archaeology and the Biblical Narrative: The Case of the United Monarchy AMIHAI MAZAR Of the various approaches to the historicity of the biblical narratives, the most justified one is in my view the claim that the so-called Deuteronomistic History preserved kernels of ancient texts and realities. This core included components of geo-political and socio-economic realia, as well as certain information on historical figures and events, although distorted and laden with later anachronisms, legends and literary forms added during the time of transmission, writing and editing of the texts and inspired by the authors theological and ideological viewpoint. The authors and redactors must have utilized early source materials, such as temple and palace libraries and archives, monumental inscriptions perhaps centuries old, oral transmissions of ancient poetry and folk stories rooted in a remote historical past, and perhaps even some earlier historiographic writings 1. This general approach to the biblical text also dictates the evaluation of the historical reality of those narratives relating to David and Solomon. The views are considerably divided: revisionist historians (the so-called minimalists ) and several archaeologists pointed out the infeasibility of the biblical description of the United Monarchy. Conservatives continue to maintain the biblical narrative as a general framework for historical reconstruction, and those who are in the middle of the road search for possible alternative historical reconstructions. 2 The 1 Cf. Miller/Hayes (1986); Halpern (1988); Na aman (1997; 2002); (2007), 399 400; Dever (2001); Liverani (2005); various papers in Williamson (2007). 2 Among the vast literature on this subject published during the last two decades I would mention the collection of essays reflecting a wide variety of views edited by Handy (1997). For conservative approaches defining the United Monarchy as a state from Dan to Beer Sheba including conquered kingdoms (Ammon, Moab, Edom) and spheres of influence in Geshur and Hamath cf. e.g. Ahlström (1993), 455 542; Meyers (1998); Lemaire (1999); Masters (2001); Stager (2003); Rainey (2006), 159 168; Kitchen (1997); Millard (1997; 2008). For a total denial of the historicity of the United Monarchy cf. e.g. Davies (1992), 67 68; others suggested a chiefdom comprising a small region around Jerusalem, cf. Knauf (1997), 81 85; Niemann (1997), 252 299;

30 Amihai Mazar archaeological paradigm concerning the United Monarchy as formulated mainly by Yadin 3 was attacked by several scholars, 4 while others continue to support this archaeological paradigm. 5 In this paper, I summarize my previous views on this subject, respond to a recent critique relating to 10th century Jerusalem, and add comments on several new archaeological discoveries relating to this subject. Summary of My Previous Views In several papers published during the last years I expressed my views concerning the United Monarchy. 6 Some of the points are summarized below (without references) and the general conclusions are cited at the end of this paper. 1. The mentioning of btdwd The house of David as a title of Judah in the Tel Dan stele, probably erected by Hazael, king of Damascus, should be given the weight it deserves. It means that about 140 years after the presumed end of David s reign, in the region David was well-known as founder of the dynasty that ruled a kingdom centered in Jerusalem. 2. The Shoshenq I raid to the Land of Israel ca. 925/920 BCE matches the mentioning of this event in 1 Kings 14:25 28. This is the only existing correlation between a biblical reference and an external written source relating to the 10th century BCE, and it means that the biblical writer must have utilized earlier documents, rooted in 10th century BCE reality. The only plausible explanation for choosing a route for this raid through the cenand Finkelstein (1999). For a middle of the road approach suggesting a United Monarchy of larger territorial scope though smaller than the biblical description cf. e.g. Miller (1997); Halpern (2001), 229 262; Liverani (2005), 92 101. The latter recently suggested a state comprising the territories of Judah and Ephraim during the time of David, that was subsequently enlarged to include areas of northern Samaria and influence areas in the Galilee and Transjordan. Na aman (1992; 1996) once accepted the basic biography of David as authentic and later rejected the United Monarchy as a state, cf. id. (2007), 401 402. For recent theoretical discussions of the emergence of the Israelite state, cf. Masters (2001); Joffe (2002). 3 Cf. Yadin (1972), 135 164, summarized in A. Mazar (1990a), 375 387. 4 Cf. Wightman (1990), Jamieson Drake (1991) and esp. Finkelstein (1996); Finkelstein/Silberman (2006); Finkelstein (2007). 5 Cf. e.g. A. Mazar (1997); Dever (1997); Meyers (1998), 243 256; Lemaire (1999), 116 120; Ben-Tor (2000); Halpern (2001), 427 478; Masters (2001); Stager (2003). 6 Cf. A. Mazar (1997; 2003; 2007a; 2007b; 2008).

Archaeology and the Biblical Narrative 31 tral hill country north of Jerusalem must have been the existence of a substantial political power in the central hill country. The most obvious candidate for such a polity is the Solomonic kingdom, and Shoshenq s goal was perhaps to terminate the rising Israelite state which threatened Egyptian economic interests. The archaeological research relating to Shoshenq I should not concentrate on looking for destruction layers in each of the sites mentioned in his list, since it is unknown whether the Egyptian army indeed violently destroyed them. Rather, the very fact that a place is mentioned in this list means that it was occupied at the time of the raid and was well-known to the Egyptians. Such an approach provides an important chronological anchor for several excavated sites throughout the country, such as Arad and Taanach, among others. The mention of Reh9ov and Beth-Shean in the list fits the archaeological evidence at those sites. 3. The list of ca. 70 names in the Negev mentioned in Shoshenq s list, some of them clearly Hebrew names, fits the unusual phenomenon of short-lived settlements known in the Negev Highlands and in the Beer-Sheba-Arad region. The material culture in these settlements represents a cultural symbiosis by the inhabitants probably people who came from Judah or the southern coastal plain who were joined by local desert nomads. The motivation for this settlement wave must have been economic, perhaps related to the contemporary large-scale copper smelting activity at Feinan (see below). The goal of Shoshenq s southern branch of his campaign was perhaps to put an end to the extensive settlement in this region, which perhaps was considered by the Egyptians as competing with or threatening their own interests. 7 4. The date of the transition from Iron I to Iron IIA is important for defining the material culture of the alleged time of the United Monarchy in the 10th century BCE (based on inner biblical chronology). The results of radiocarbon dates relating to this transition can be interpreted in various ways: while Sharon et 7 The concept of a Tel Masos Chiefdom centered at Tel Masos and including the Negev Highland sites, as suggested by Finkelstein, is highly questionable. Tel Masos is located in a different geographic zone (Arad-Beer-Sheba valley) than the Negev Highland sites, its ceramic repertoire seems to be earlier than that of the Negev Highland sites and it lacks the hand-made pottery (probably produced by local nomads) which comprises about 50% of the pottery in the Negev Highland sites.

32 Amihai Mazar al. insist on dating the transition to ca. 900 BCE, 8 Finkelstein, who since 1996 dated the transition to Shoshenq s time, now corrected his view (at least in relation to the end of Megiddo VIA) and claims an earlier date in the 10th century BCE for that violent destruction, which marks the end of the Iron Age I at Megiddo. 9 Utilizing the data published by Sharon et al., Bronk Ramsey and myself calculated that the transition must have occurred during the first half of the 10th century BCE, which would fit with Finkelstein s recent view. 10 This enables us to determine the alleged date of the archaeological evidence related to the United Monarchy to the transition of Iron I/IIA and to the early part of Iron IIA. 11 5. Demographic assessments of 10th century BCE Judah are questionable, since they are based on surface surveys of sites which in many cases were settled continuously for most of the Iron Age. Both temporal and spatial aspects of the development of such sites remain enigmatic in such surveys, and thus calculations of the numbers of sites and the settled areas during the 10th and 9th centuries BCE are susceptible to significant errors. In spite of these limitations, the comparison of the population estimation in Iron I (based on excavations and surveys) to that in the late 8th century BCE enables to presume a gradual increase in population throughout this time duration. A population estimation of about 20,000 people for all of Judah and Benjamin in the Iron IIA (including the Shephelah) seems to be possible, though the methodological difficulties mentioned above should be taken into account. This number, if correct, provides a sufficient demographic basis for an Israelite state in the 10th century BCE. 6. Revival of urban life following demise of urbanism in large parts of the country during the Iron Age I is detected in excavated sites throughout the Israelite territories from Galilee to Judah. This was a gradual process which continued until the late 8th century BCE. Many of the sites remained unfortified and not sufficiently developed as urban centers during the 10th century, while others were fortified (see below). Revival of trade with Cyprus occurred during the Iron IIA. 8 Cf. Sharon et al. (2007; 2008). 9 Cf. Finkelstein/Silberman (2006), 180 182. 10 Cf. A. Mazar/Bronk Ramsey (2008); A. Mazar (2008), 100 105, 112 115. 11 Cf. A. Mazar (2007a; 2008).

Archaeology and the Biblical Narrative 33 7. Tel Reh9ov in the Beth-Shean Valley demonstrates continuity of a large 10 ha city throughout the 12th 9th centuries BCE. Yet, while during the Iron Age I (12th 11th centuries BCE), Canaanite material culture is dominant, the 10th century BCE (Iron IIA) sees a considerable change in the material culture (mainly the appearance red-slipped and hand burnished pottery). This change can be detected in many other parts of the country at almost the same time, and may be regarded as reflecting geopolitical developments that took place during the 10th century BCE, perhaps related to the emergence of the Israelite state. 8. Yadin s identification of Solomonic cities at Hazor (Stratum X), Megiddo (Stratum IVB VA) and Gezer (Stratum VIII), thus illuminating 1 Kings 9:15, is still a debated subject. Finkelstein and his followers abandon this theory altogether, yet the current excavators of Hazor and Gezer support Yadin s theory. The new excavations at Megiddo provided two relevant 14 C dates from Level H-5, which corresponds to Stratum IVB VA: one in the 10th century and the other in the 9th century BCE. Dates from the destruction of Megiddo VIA fit the late 11th or early 10th century BCE. 12 These dates suggest that Stratum IVB VA, with its two ashlar palaces, could have been constructed during the 10th century BCE and thus could have been Solomonic, although additional radiometric dates are required. 9. The discovery of inscriptions with the name Hanan at Beth- Shemesh and Timnah (Tel Batash) along the Sorek Valley in Iron IIA contexts recall the name Elon Beth Hanan among the places in Solomon s second district, mentioned in 1 Kings 4:9. This adds support to the possible 10th century origin of this biblical administrative list. 10. The small amount of Hebrew epigraphic finds from the 10th century BCE was brought as evidence for lack of literacy during the 10th century and thus for the infeasibility of an Israelite state during this century. However, the number of Hebrew inscriptions from Israel in the 9th century is also very small, and yet there is no debate concerning the existence of an Israelite state in that century. New finds from Tel Zayit and Khirbet Qeiyafa (see below), as well as those mentioned in the previous paragraph, may indicate that during 10th century literacy in Judah was much more advanced than presumed in earlier studies. 12 Cf. A. Mazar/Bronk Ramsey (2008); A. Mazar (2008).

34 Amihai Mazar Questions related to Jerusalem and several new discoveries are the subjects of the following part of this article. Jerusalem in the 10th Century BCE The status of Jerusalem as a city in the 10th 9th centuries BCE has become a major subject of debate. While in the past, archaeological assessment of the United Monarchy tended to ignore the problems concerning Jerusalem, some current authors use the Archaeology of Jerusalem as a major issue in deconstructing the historicity of the United Monarchy. Thus, Ussishkin claimed that Jerusalem was not settled in the 10th century and Finkelstein defined 10th century Jerusalem as a small village. 13 The topography of Jerusalem indeed does not allow to recreate a very large city there prior to its extension to the western Hill during the 8th century BCE. The eastern ridge of the City of David and the Temple Mount comprise about 12 ha, and excluding the temple mount the area is just 4 5 ha. Such a city could not include a population larger than ca. 1000 2000 persons, and such a small city can hardly be imagined as a capital of a large state like the one described in the Bible. However, several exceptional structures that were excavated in this city set it apart from other urban centers of the southern Levant at that time. These include the architectural complex on the summit of the City of David, the possible continued use of the Middle Bronze structures around the spring Gihon, and the temple, known only from biblical descriptions. These real and virtual structures, if correctly dated and understood, may throw light on the power base for rulers such as David and Solomon, providing that we correctly define the nature of their kingship and state. The Stepped Structure and the Large Stone Structure Complex The Stepped Structure in Shiloh s Area G and the Large Stone Structure excavated by Eilat Mazar to its west, should be defined as part of one and the same architectural complex. 14 Each of the three excavators of these buildings (Kenyon, Shiloh and E. Mazar) dated them to the 13 Cf. Ussishkin (2003); Finkelstein (2003). 14 Cf. E. Mazar (2008).

Archaeology and the Biblical Narrative 35 Iron I or Iron IIA and related them to the United Monarchy. 15 This date and interpretation were recently challenged by Finkelstein, Ussishkin, Herzog and Avitz-Singer. The importance of this debate for our subject calls for a detailed response, which is the subject of the following paragraphs. 16 The Stepped Structure 17 Various parts of the Stepped Structure in the City of David (Fig. 1) were exposed by Macalister, Kenyon and Shiloh, and the excavation of its northern face was recently accomplished by E. Mazar. 18 This is a large structure, about 40 48 m long and ca. 20 m high. 19 It includes several components, the most prominent being the mantle wall, a term used by Cahill to describe the outer sloping stepped structure, which in her view was founded on a massive substructure denoted by Kenyon and Shiloh as terraces. The latter are explained as a constructional feature, creating stone boxes filled with stones and intended to support the mantle wall on the steep slope of the hill. In certain places, there are earth layers between the stone terraces and the mantle wall, but this is not consistent and in other places the mantle wall was constructed right on top of the stone substructure or, in fact, is bonded to it. 15 The terraces below the Stepped Structure were dated by Kenyon (1974) and Shiloh (1984) to the Late Bronze Age, yet they were redated by Steiner (2001) and Cahill (2003) to Iron Age I and defined as the substructure of the Stepped Structure, based on a room containing Iron Age I pottery found by Kenyon below the terraces, and the Iron I pottery found inside those terraces. 16 The discussion below refers to Finkelstein et al. (2007). My thanks to Eilat Mazar for guiding me several times in her excavation areas during the 2007 and 2008 seasons and discussing with me some of the issues raised in the following discussion. Yet, the views in the following response are mine. 17 This building is usually called The Stone Stepped Structure. Here it is abbreviated to The Stepped Structure. 18 Cf. Shiloh (1984), 15 17; Steiner (2001), 36 39, 43 48, 51 53; Cahill (2003); E. Mazar (2007a; 2007b; 2008). 19 The height of 27.5 m of this structure cited by E. Mazar (2008), 30, is based on including the Large Wall in Kenyon s Trench I as part of the Stepped Structure. Though this is the view of Steiner (2001) and Cahill (2003) as well, I am not confident that this wide wall was part of the same complex (see below). The width of 48 m cited by E. Mazar (op. cit.) is based on adding structural remains exposed by Macalister/Duncan (1926) south of the Hasmonean Tower.

36 Amihai Mazar Fig. 1: The remains of the Stepped Structure and the Large Stone Building complex as revealed by the excavations of K. Kenyon, Y. Shiloh and E. Mazar. Component 1: Component 2: Component 3: Component 4: Component 5: Component 6: The terraces (structural foundations of the Stepped Structure ) The mantle wall of the Stepped Structure A stone structure or fill (probably part of the Stepped Structure ) in Kenyon s Square AXXIII Terraces 4 5 in the upper part of Kenyon s Trench I The Large Wall in the upper part of Kenyon s Trench I The Large Stone Structure excavated by E. Mazar Combined plan based on plans published by Shiloh [1984], Steiner [2001] and E. Mazar [2009a: 38 Fig. 1; 2009b: 64]. Computer work by Y. Shalev.

Archaeology and the Biblical Narrative 37 The following is a list of points raised by Finkelstein et al. concerning this Stepped Structure and the corresponding responses. 20 1. Finkelstein et al. suggest that the Stepped Structure had two building phases. Its lower part is a later addition, since it was constructed of smaller stones. 21 The stones in the lower 17 courses are indeed 0.20 0.35 m in size while those in the upper 35 courses are 0.35 0.7 m long (a few are up to 1 m long), yet this difference is just a technical matter; the lowest course of large stones was constructed just above the highest course of smaller stones and thus the former could not predate the latter. There is no evidence for two construction phases, and both parts are superimposed by Iron Age II dwellings. The reason for the change in stone size is perhaps related to the challenge faced by the builders when they approached the steep vertical rock scarp behind the upper part of the structure. 22 The purpose of the Stepped Structure was probably to support the foundations of a large building constructed on top of the hill by covering the vertical natural scarp with its inner cavities and karstic features and extending the area to the east. The change in orientation between the lower and upper parts is mentioned by Finkelstein et al. as additional evidence for two construction phases. Yet, this change is gradual: The lower courses of large stones follow the same orientation as the courses of the smaller lower stones, and as we proceed upwards the courses start to turn to the northwest, in accordance with the topography. Thus, the suggestion for two construction phases is intangible. 2. The authors cite Steiner s mention of Iron IIA pottery among the stones of Components 3, 4, 5 23 and suggest (although with reservation) that this pottery provides a terminus post quem for the construction of the Stepped Structure. As I have shown elsewhere, this pottery came from unclear contexts above or between the upper stones of Component 5 ( The Large Wall ) in Kenyon s Trench I. 24 No floor or any other occupation layer related to this wall was ever excavated. I claimed (and Finkelstein et al. agreed) that since Components 4, 5 in Kenyon s Trench I are detached from the main part of the Stepped Structure, 20 Cf. Finkelstein et al. (2007), 142 164. 21 Cf. Finkelstein et al. (2007), 151. 22 The latter was clearly revealed by E. Mazar in the 2007 2008 excavation seasons. 23 Op. cit., above n. 20. The numbers refer to Fig. 1 in this paper. 24 Cf. A. Mazar (2006), 263 264.

38 Amihai Mazar there is no proof to Steiner s claim (accepted also by Cahill and E. Mazar) that they were part of this structure. In addition, the above mentioned pottery group includes only a few pottery sherds, mostly dating to Iron I but a few undefined sherds. A single almost complete vessel is probably of Iron IIA date, but as said above, it has no chronological value in establishing the date of either the Stepped Structure or even of the The Large Wall itself. 25 3. Cahill published Iron IIA pottery, including an imported Phoenician Bichrome jug, found on the earliest floor surfaces of the Burnt Building above the lower northern part of the Stepped Structure. 26 According to Cahill, this pottery provides a terminus ante quem in the Iron IIA for the construction of the Stepped Structure. Finkelstein et al. claim that the floor surfaces were in fact constructional fills for the late Iron II building. 27 I prefer the interpretation of the excavators as presented by Cahill. If the layers were constructional fills laid in a later period, we would expect some mixture of pottery, and yet these layers contained purely Iron IIA pottery. Even if they were constructional fills, they must have been constructed no later than Iron IIA and thus substantiate the terminus ante quem for the construction of the Stepped Structure. 4. The authors claim that the upper part of the Stepped Structure is a rebuild of the Hellenistic period or even a modern reconstruction. 28 As to the latter claim, modern reconstructions were indeed made by the Jordanian authorities before 1967 near the northern corner of the Great Tower of the Second Temple First Wall south of Shiloh s Area G, but not in the latter area, except for some reinforcement with cement of several existing stone courses. 29 As to the former claim, the Second Temple period city wall (Shiloh s Wall 309, E. Mazar s Wall 27) was indeed constructed just above the upper part of the Stepped Structure (E. Mazar s Wall 20) and at places it joined the latter where it was well-preserved. This can be seen, for example, in the southern part of Area G, where the Second Temple period wall continues from Macalister s Great Tower (Shiloh s Wall 310) until the 25 Cf. A. Mazar (2006) for discussion and references. 26 Cf. Cahill (2003), 56 66. 27 Cf. Finkelstein et al. (2007), 152. 28 Cf. Finkelstein et al. (2007), 152 155. 29 Cf. Shiloh (1984), 62, Fig. 27 shows the reconstruction at the corner of Walls 310 and 309.

Archaeology and the Biblical Narrative 39 upper part of the Stepped Stone Structure, until it joined the Northern Tower (Shiloh s Wall 308). 30 In his Squares C 1 2, Shiloh excavated the top of the Stepped Structure, indicating that the wall was at least 5 m wide, though he did not reach its western face. 31 It was clear to him that this wide wall was the upper part of the Stepped Structure and that it preceded Wall 309 of the Second Temple period. This was further clarified by E. Mazar s excavations: her Wall 20 (which is, in fact, the upper part of Shiloh s Wall 302) was exposed in sections along a total length of 22 m; its width was 5.8 m and its western face was preserved to a height of 1 1.8 m. 32 A 0.8 m thick layer of Iron I occupation debris abutted the western face of Wall 20 at the southern end of the excavation area. Both Kenyon and Shiloh found remains of an earth glacis dated to the Hellenistic period which covered the Stepped Structure and abutted the Second Temple period wall, creating a support for this wall against erosion on the steep slope. 33 Finkelstein et al. s suggestion that both, the upper part of the Stepped Structure as well as the glacis, were part of a single building project of the Hasmonean era contradicts the facts: These are two different building projects, each with its own function. During 2006 and 2007, E. Mazar dismantled a part of the Northern Tower of the Second Temple period (Shiloh s Wall 308) and found that it was built against the earlier Wall 20 of the Stepped Structure, and its upper part relates to the Second Temple period (Shiloh s Walls 309 equal to E. Mazar s Wall 27). 34 Wall 20 was founded on a rock 30 Cf. Shiloh (1984), 62, Fig. 27; also ibid., 55 Fig. 17; and the photos and drawing in E. Mazar (2009a), 24 25, 27 28; id. (2009b), 37, 58. On the photos one can see how Wall 309 (= E. Mazar s Wall 27), the city wall of the Second Temple period, is founded on the upper part of the Stepped Structure (Shiloh s Wall 320 = E. Mazar s Wall 20). In the southern part of Area G, north of the Southern Tower, the Second Temple Wall 309, preserved 6 7 courses high, abuts the mantle wall of the Stepped Structure which in this place was preserved until the present topsoil, at the same level as the 7th course of the Second Temple Wall. In E. Mazar s excavations the separation between Wall 20 (the upper part of the Stepped Structure ) and Wall 27 (the Second Temple city wall) became clear; there is a slight difference in their orientation, though they were constructed one on top of the other. 31 Cf. Shiloh (1984), 55 56; Figs. 16 17. The upper part of Wall 302 corresponds to E. Mazar s Wall 20. 32 Cf. E. Mazar (2007b), 15, 21 Fig. 1, 24 Fig. 5; id. (2009b), 56. For isometric drawing cf. E. Mazar (2009a), 28; id. (2009b), 65. 33 Cf. Shiloh (1984), 30, 55 Fig. 17. 34 Cf. E. Mazar (2007a), 71 75, plan on p. 73 and photograph on p. 87, lowest end, and also id. (2009b), 72 79.

40 Amihai Mazar scarp and was clearly bonded with the upper courses of the Stepped Structure. Wall 20 and the rock scarp on which it was founded was abutted on its eastern face by thick debris layers; the upper ones contained early Persian/Babylonian Period pottery and other finds, while the lower ones contained rich deposit of finds from the end of the Iron Age, among them several dozens of fragments of inscribed clay bullae. This layer appears to have been dumped from a building to the west, apparently the Large Stone Structure which stood at higher elevation (see below). From a structural point of view, there is no doubt that Wall 20 and the Stepped Structure are contemporary. Wall 20 cannot be dated to the Hellenistic period as argued by Finkelstein et al. The Large Stone Structure The Large Stone Structure is a term given by E. Mazar to a building which she excavated on the summit of the hill west and northwest of the Stepped Structure (see Fig. 1). 35 Its walls are 2 5 m wide, its width was at least 30 m, and its length is unknown. Since only a few walls and segments of floors of this structure were preserved, and the area was much disturbed by Herodian and later activity, as well as by Duncan and Macalister s excavations, the deciphering of its architecture and date are not a simple task, as explained by E. Mazar in her preliminary publications. Finkelstein et al. present a wholesale denial of the excavator s interpretation of the plan, nature and date of this building. In the following, I will examine their arguments. 1. As explained in the previous section, Wall 20, the eastern wall of the Large Stone Structure, is also the upper part of the Stepped Structure and thus cannot be later to this structure, as suggested by Finkelstein et al. 2. The earth layer found above bedrock and below the walls of the Large Stone Structure, contained Iron I pottery (as well as Middle Bronze and some Late Bronze sherds). Finkelstein et al. claim that this layer should not be considered when dating the construction of the building. 36 Indeed, in principle, pottery found in earth layers below foundations of buildings can provide just a terminus post quem for the construction of the build- 35 Cf. E. Mazar (2007a), and also id. (2009b), 43 65. 36 Cf. Finkelstein et al. (2007), 147 148.

Archaeology and the Biblical Narrative 41 ing above. It should be recalled, however, that establishing a foundation date for an excavated building is a difficult task in most cases. While finds found on floor surfaces provide a date for the final use of a building or to the longevity of its use, its foundation date is always enigmatic, and depends to a large extent on the finds in earlier occupation levels, foundation trenches, constructional fills, etc. Kenyon, for example, argued that it is commonplace in British archaeology that a building is dated by the latest object in its building deposits, i.e. foundation trenches, floor make-up and so on. 37 This argument certainly cannot be taken as a general rule, and there are numerous variations: each case should be judged independently. In our case, both Kenyon and Shiloh found that the latest pottery in the constructional fills of the Stepped Structure was Iron Age I and E. Mazar found the same pottery assemblage in the earth layer below the Large Stone Structure. This earth layer abutted the lower parts of the foundation stones of the building and fragmentary floors of the building were found just above this layer. If the Stepped Structure and the Large Stone Structure were constructed at a later date, we would expect to find at least a few post-iron I sherds in these layers, yet, this is not the case. Since the two structures are bonded (as indicated by Wall 20) and the pottery found by three expeditions in all the constructional fills and layers below the foundations is homogeneous and uncontaminated, it is justified in my view to claim that the Iron I pottery is as close as it can be to the construction date of this large architectural complex. 3. Finkelstein et al. claim that the pottery assemblage in the abovementioned earth layer is as late as 10th 9th century BCE. However, as mentioned above, this pottery is identical to that found in the constructional layers and foundation terraces of the Stepped Structure 38 and it is similar to Iron Age I contexts at sites like Giloh (12th century BCE) and Shiloh Stratum V (11th century BCE). 39 37 Kenyon (1964), 145. 38 Cf. Steiner (2001), 36 39, 43 48; Cahill (2003), 46 51. 39 The best parallels to the cooking pots from the earth layer are those from Shiloh: Finkelstein et al. (1993), Fig. 6.47:1 5 on p. 165 and Fig. 6.50:1 2 on p. 169 dated by the excavators to the 11th century BCE (ibid., 163, 168). The argument of Finkelstein et al. (2007), 148, that there was at least one rim which seems to date to the late Iron I or early Iron IIA was based on an impression from a single visit to the site and from a single photograph of rim sherds. However, the drawings published by E. Mazar (2007a), 50, include only Iron I sherds. Several cooking pot rims have a

42 Amihai Mazar 4. Finkelstein et al. argue that Wall 107, the main wall of the Large Stone Structure, should be divided into an eastern part and a western part, each belonging to a separate structure. 40 Indeed, there is a slight difference in orientation between the eastern and western parts of the wall, yet, this could be due to topographic constraints. The gap between these two parts of the wall was caused by the foundations of a Second Temple vaulted underground room (see below). Although Wall 107 was badly preserved, and most of its southern face is missing, the construction technique of the eastern and western parts are similar, and both were founded above the same earth layer containing Iron I and earlier pottery. On its eastern end, Wall 107 creates a corner with Wall 20. 41 Since the latter served as the western wall of the Stepped Structure (see above), as well as the eastern wall of the Large Stone Structure, the two must be contemporary and belong to the same architectural complex. Other walls which corner with Wall 107 (Walls 19, 21, 109) must be a part of the same complex as well. 5. Finkelstein et al. claim that the eastern part of Wall 107 should be dated to the Hellenistic period, since stones of this wall are seen in a photograph above the eastern wall of a vaulted chamber (Walls 69, 72, 71) of the Second Temple period. 42 This argument is flawed, since the chamber was clearly later than Wall 107. The picture was taken after the removal of plaster and other parts of the vaulted chamber. The builders of this Second Temple period underground room left large stones of Wall 107 in place wherever it was not necessary to remove them, utilized these stones as part of their new construction and covered them by plaster. Such plaster was never used in other parts of the Large Stone Structure. 6. Finkelstein et al. claim that a ritual bath (a miqweh; Walls 61, 63, 66) should be regarded as belonging to the eastern part of the Large Stone Structure and thus the two should be dated to the Second Temple period. 43 However, this ritual bath is one of sevmolded rounded everted rim, such as E. Mazar (2007a), 50, nos. 12 14. Only a few rims have a concave depression along the outside of the rim such as E. Mazar (2007a), 50, no. 11. However, in all these cases, the rims are everted, indicating an early date in Iron Age I. 40 Cf. Finkelstein et al. (2007), 155 157. 41 Cf. E. Mazar (2007a), 59; plan and photo on p. 87, reproduced in Finkelstein et al. (2007), 158, Fig. 5. 42 Cf. Finkelstein et al. (2007), 154 157. For the photograph see E. Mazar (2007a), 74. 43 Cf. Finkelstein et al. (2007), 154 157.

Archaeology and the Biblical Narrative 43 eral such baths, cisterns and pools, dated to various periods (from the Second Temple period until the Islamic period) which penetrated into the excavation area from higher occupation levels. Finkelstein et al. claim that the bath was part of the Large Stone Structure is based just on its orientation. Yet, even this claim is incorrect: the western wall of the bath (Wall 66) runs on an angle compared to Wall 67 of the Large Stone Structure (unlike in their flawed reconstructed plan). 44 Like in the case of the vaulted chamber, the building technique of the ritual bath differs completely from that of the Large Stone Structure : while the former was constructed with plaster typical of Second Temple architecture, the latter was constructed of large, roughly cut stones without the use of plaster. 7. Finkelstein et al. claim that the Iron IIA pottery assemblage published by E. Mazar from Locus 47 in Room C of the Large Stone Structure has no significance, since it was not found on a floor and since it contained also Iron IIB pottery. 45 However, although this pottery group was not on a floor, it was found as a homogeneous deposit, including a few restorable vessels and large sherds of typical Iron IIA horizon, located in a very small space which was enclosed on all four sides by massive walls: Walls 19 and 21 (both abutting Wall 107) and the subsidiary (though massive) Walls 22 and 24. These walls were preserved to a height of 1.2 1.4 m, and the pottery was found close to their lower parts. It is plausible that this pottery was slightly moved from its original place when Walls 22, 24 were added, yet, the group retained its nature as a homogeneous, partly restorable, assemblage. 46 8. The continuation of Walls 19 and 21 of the Large Stone Structure was found by Kenyon in her Area H1, just a few meters to 44 Cf. Finkelstein et al. (2007), 158 196, Figs. 5 6. 45 Cf. Finkelstein et al. (2007), 149. The assemblage was published in E. Mazar (2007a), 66, with the photo on the left on p. 63. 46 The argument of Finkelstein et al. (2007), 149, that the lower part of Locus 47 contained Iron IIB pottery is based on the basket number of a single bowl rim sherd: E. Mazar (2007a), 70 sherd no. 7. The authors argue that since this basket number is in the same range as the basket numbers of the Iron IIA cache, it must have originated from the same context. Yet, a basket number is just a technical device, and the sherd might have come from an upper level of this locus, regardless of the basket number. E. Mazar (2009a), 37; id. (2009b), 66, argues that the Large Stone Structure continued to be in use until the end of the Iron Age. During its use, changes were made in the building, as evidenced by the additions of walls like Walls 22 and 24 on both sides of the Iron IIA pottery cache. Few Iron IIB sherds could penetrate to lower levels during such building operations.

44 Amihai Mazar the north of this building (her Walls 91 and 92, each 2 m wide with 1.3 m space between them; see Fig. 1). 47 Kenyon dated these walls to the 10th century BCE and Steiner writes that although no pottery was found on the plaster floor of the structure, there were 10th 9th centuries BCE sherds in the fill above the floor between the two walls. The pottery from this trench was never published in drawings, but we may suppose that Kenyon and Steiner s dating was based on red-slipped and hand burnished vessels, known to them as typical of the 10th century BCE. 9. Finkelstein et al. argue that Iron IIA pottery was found below architectural elements in Room B (west of Locus 47) and thus the Large Stone Structure must be later than the Iron IIA. Yet, E. Mazar wrote that the Iron IIA pottery from this room was found below a bench and a stone pavement, which are attributed to later phases of the building. 10. E. Mazar claims that the Large Stone Structure continued to be in use until the Babylonian conquest of Jerusalem. In fact, very few Iron II remains were found in the excavations, all in disturbed layers or between the collapse of the upper stones of the structure. The contexts of Iron II finds revealed by Macalister and Duncan and cited by Finkelstein et al. are unknown. As the authors admit, Herodian pottery sherds in the stone debris may have infiltrated either during Herodian activity in the area or by Macalister and Duncan s excavations. 11. Finally, Finkelstein et al. published two suggested reconstruction plans of the Large Stone Structure, which they attribute to the Second Temple period. 48 The architectural elements in this reconstructed plan belong, in fact, to three different periods: the Iron Age, the Second Temple and the Byzantine period. As explained above, Walls 20 and 107 must be Iron I or Iron IIA, at the latest. Walls 21 and 19 are perhaps an Iron II addition. The ritual bath (Walls 61, 63, 66) is from the Second Temple period, and the southern wall termed in the drawing as Inner Wall is Macalister s Davidic Wall ; it was exposed by E. Mazar during 2008 and dated to the Byzantine period. Thus, this reconstructed plan should be dismissed. 47 Cf. Kenyon (1974), 115 and the photograph on p. 37; Steiner (2001), 48 49; reconstruction in E. Mazar (2007b), 24 Fig. 5, right side. See Fig. 1, walls to the right of Walls 19, 21, 23 and 24. 48 See n. 39 above.

Archaeology and the Biblical Narrative 45 In light of the above, the archaeological arguments presented by Finkelstein et al. are unacceptable. The Stepped Structure and Large Stone Structure should be seen as one large and substantial architectural complex. The former must be explained as a support structure of the latter, which stood on the summit of the ridge to the west, on the narrowest point of the City of David spur, which was naturally bounded by an almost vertical rock cliff on the east. Cahill claimed that the construction date of the Stepped Structure must have been either contemporary or shortly later than the pottery found in its substructure, which is clearly Iron Age I in date, while Kenyon, Shiloh and Steiner suggested a 10th century BCE date for its construction. 49 The same argumentation is valid for the Large Stone Structure. The magnitude and uniqueness of the combined Stepped Structure and the Large Stone Structure are unparalleled anywhere in the Levant between the 12th and early 9th centuries BCE. Shiloh suggested that the Stepped Structure was intended to serve as a substructure for the upper structure of the citadel of the City of David, built there over the remains of the Jebusite citadel. 50 E. Mazar suggested that the Canaanite citadel was further to the south (in an unexcavated area), and that the Stepped Structure and Large Stone Structure complex should be interpreted as David s palace, i.e. were constructed during the early 10th century BCE. I suggested to identify the entire complex with Metsudat Zion the fortress of Zion mentioned in the biblical description of David s conquest of Jerusalem. David is said to have changed the name of this citadel to `Ir David, the city of David (2 Sam. 5:7, 9). 51 This identification is suggested with due caution, since it is based on two rather shaky pillars: the one is the possible Iron Age I construction date of the entire complex. The other is the above mentioned biblical text, the historicity of which may be questioned. We should also note that the Jebusites, the supposed builders of this citadel, are unknown to us from any sources outside the bible, and Archaeology did not provide any particular characteristics of such an independent ethnic group. 52 Finkelstein et al. conclude their paper with 49 Cf. Cahill (2003), followed by A. Mazar (2006). 50 Cf. Shiloh (1984), 17. 51 Cf. A. Mazar (2006), 265. 52 At Giloh, a small Iron I site 7 km southwest of the City of David, I uncovered the remains of a massive square structure dated to the Iron Age I (probably 12th century BCE) which I thought to be a foundation of a tower (Mazar 1990b). The massive structure and its building technique recalls to some extent the large substructure of the Stepped Structure. I identified the site as early Israelite while Ahlström (1984) suggested to identify it as a Jebusite site. The pottery from Giloh resembles the as-

46 Amihai Mazar an admonition against such straightforward identifications of structures mentioned in biblical texts which were written much later. Yet, as mentioned in the beginning of this paper, the historicity of the biblical narratives and the relationship between text and Archaeology are subject of continuous debate. There is no absolute truth in this field and we must accommodate pluralism and a wide spectrum of views. I agree with Finkelstein that objective archaeological criteria are essential for examining biblical narratives whenever this is possible. Many scholars argue that the so-called Deuteronomistic History, as well as other biblical sources, preserved old memories and knowledge of the past to a certain degree, although these could have been distorted during transmission and editing processes, as noted in the beginning of this paper. In the case of Jerusalem, the preservation and transmission of historical memories during hundreds of years is a feasible possibility, since the city did not suffer from any turmoil between the 10th and 7th centuries BCE. Old inscriptions and other written texts, as well as oral transmission of information, could be preserved over centuries. Finkelstein argued that David s biography as a young leader of a warrior gang is historical, since, in his view, the narrative fits the archaeological background relating to the late Iron I. However, he denies David s biography as a king, since, again in his view, it contradicts the archaeological picture of the 10th century BCE in general, and that of Jerusalem in particular. 53 However, if the Iron Age I or Iron IIA date of the citadel complex (the Stepped Structure and the Large Stone Structure ) is accepted, then the archaeological profile of Jerusalem before or during the presumed time of David would be very different from that presented by Finkelstein and Ussishkin. Such a profile shows that Jerusalem was a rather small town with a mighty citadel, which could have been a center of a substantial regional polity. 54 semblage found in the substructure of the Stepped Structure and Large Stone Building in the City of David. 53 Cf. Finkelstein (2003), 89, 91. 54 Cf. Finkelstein (2003); Ussishkin (2003) against A. Mazar (2007a), 152 154.

Archaeology and the Biblical Narrative 47 Additional Discoveries in the City of David Iron IIA pottery was found in all of the areas excavated by Shiloh on the eastern slope of the City of David. 55 According to the Modified Conventional Chronology which I and many others utilize this pottery may be dated to the 10th 9th centuries BCE, while a more precise distinction needs further research. 56 The fact that almost no Iron IIA architecture was preserved on the eastern slope of the City of David should probably be explained as a result of erosion, the continued use of stone structures over hundreds of years, the robbing of older building materials by later builders, and rock quarrying, all of which caused a distortion of the archaeological picture in Jerusalem. The lack of Late Bronze structures should be explained along the same line, and clearly stands in contrast to the information gained from the Amarna letters from Jerusalem. 57 Discoveries made by Reich and Shukron in their excavation at the Gihon spring during the last fifteen years include massive structures around and west of the spring that were probably part of a large fortified citadel, a large quarried space in the rock dubbed a pool, and the cut of the original (upper level) tunnel known as part of Warren s Shaft. 58 These components were dated by the excavators to the Middle Bronze Age. The fortifications are among the mightiest ever found in any Bronze or Iron Age site in the southern Levant, and thus they are evidence for a central powerful authority and the outstanding status of Jerusalem during the Middle Bronze Age. This special status might have been retained in the local memory until the end of the second millennium BCE and later, and perhaps is one of the main reasons for the choice of Jerusalem as a capital of the newly established kingdom during the Iron Age. We have to ask whether this magnificent architectural system went out of use by the end of the Middle Bronze Age. New discoveries, made in 2008 by Reich and Shukron, have shown that 55 For Iron IIA pottery from Shiloh s excavation cf. de Groot/Ariel (2000), 35 42, 93 94, 113 121, Figs. 11 15. The pottery from Area E will be published in a forthcoming volume of Qedem submitted by A. de Groot and H. Greenberg. Iron IIA pottery from Area G was published by Cahill (2003), 59 62. 56 Cf. A. Mazar (2005). Herzog/Singer-Avitz (2004) suggested inner division of the period into an early and late sub-periods, dated to the 10th and 9th centuries BCE accordingly. Yet, the attribution of the assemblage from Jerusalem to one of these periods is still unclear. The substantial finds from this period in Jerusalem excludes their suggestion that Judah emerged as a state in the southern Shephelah and the northern Negev rather than in the hill country. 57 Cf. Na aman (1996); Millard (2008). 58 Cf. Reich/Shukron (2008).

48 Amihai Mazar the two east west massive walls (about 5 m wide) of the tower west of the Gihon spring continued westwards up the slope until they joined the bedrock scarp close to the horizontal tunnel of Warren s Shaft. The northernmost of these two walls, constructed of incredibly large stones, still stands to a height of over 8 m! 59 During the Iron Age II, this system was well-known, as can be learned from three features: 1. Late Iron Age II walls abut walls of the Middle Bronze fortification system at several points. 2. During the Iron Age IIA (9th century according to the excavators), the large rock-cut area (so called pool ) south of the abovementioned tower was well-known, since it was entirely filled with earth and large stones that served as a constructional fill for an Iron Age II building. This fill contained over 180 unepigraphic seal impressions on bullae dated to the 9th century BCE, as well as thousands of fish bones. 60 3. The deepening of the Warren s Shaft system and the discovery of the natural karstic shaft occurred, according to Reich and Shukron, sometime during the Iron Age II, but before Hezekiah s tunnel was cut in the 8th century BCE. This indicates that the original upper part of the system was known and probably in use since the Middle Bronze Age through the 9th century BCE. 61 It thus may be suggested that the immense Middle Bronze fortifications and pool were also in continuous use until the Iron Age II, although there is no actual ceramic or other direct proof for this longevity, perhaps due to continued cleaning and renovations of this area throughout this long period. As to the Temple Mount, if it was indeed part of the city during the time of Solomon, it more than doubled the area of Jerusalem to ca. 12 ha. This new area could provide plenty of space for public buildings as those described in the biblical texts: Temple and palace, and perhaps elite residencies. Yet, the answer to the question whether such buildings indeed stood in Jerusalem during the 10th century BCE depends on one s approach to the biblical text, as no direct archaeological evidence is available. In an earlier discussion of this issue, I asked the question: if Solomon did not built a temple in Jerusalem, who was responsible for the construction of the Jerusalem temple later in the Iron Age? 62 The architectural parallels between the biblical description of the Jerusalem temple to north Syrian temples, like those at Tel Taynat and 59 I thank R. Reich and E. Shukron for showing me their recent discoveries. 60 Cf. Reich/Shukron/Lernau (2007). 61 This was already suggested by Cahill (2003). Recall that Kenyon suggested such a continuity in relation to the much scantier Middle Bronze wall which she found higher on the slope. 62 Cf. A. Mazar (2007a), 154; Liverani (2005), 329, who is skeptical concerning the validity of the biblical description, yet, does not exclude a modest Solomonic temple.